Issue No.:HB20181001E Date:2018-10-01 Author:Chang I-Hua, Hsu Cheng-Mei, Wu Chuan-Wei More News facebook

Change language>>简体中文 繁體中文

Teaching Behavioral Data and Flipped Learning

Chang I-Hua Professor, Chair
Graduate Institute of Educational Administration and Policy, National Chengchi University
Hsu Cheng-Mei Assistant Professor
Department of Visual Communication and Design, China University of Technology
Wu Chuang-Wei Honorary President
Chang I-Hua Professor, Chair
Taiwan Technology Leadership and Instructional Technology Development Association

I. Introduction

Flipped Classroom, originating in the USA, 2007, is a new teaching model. The model demands students to review teaching materials prepared by teachers or others at home before working on assignments and conducting discussion on questions with classmates and teachers in class. The model differs from traditional methods in that students switch their role with teachers, learning at home and completing assignments in class (Wikipedia, 2018). In the digital era, limited teaching time and material capacity should no longer affect the way students learn and teachers teach, being problems of the past. Therefore, teachers, who are responsible for education, can no longer teach with ways they “learned” in the past.

Students “nowadays” are to adapt themselves to the “future” (Wu, Yao-Ming, 2010) However, with the constant changes of emerging technologies, students of digital-native generations need educators who upgrade over time and advance their technology proficiency so that students can be prepared for a successful future in an even more technological era (Brooks-Young, 2006). Emerging technologies are changing our education. Emerging technologies such as Augmented reality (AR), 3D printing, cloud computing, online social networking, flexible display, biometrics, multi-touch screens and game-based learning (GBL) are now being used or will be used in the classroom in the future(Hongkiat, 2017). Various future classrooms which consist of aforementioned components are influencing the way teachers and students utilize technologies across the world as well as innovations and changes of teaching scenes within our country. Thus, teachers aiming to innovate their teaching should start from now on. Professional teachers should use “innovative” teaching models to teach students “nowadays” to adapt to “future” life. Teachers should beware not to fall behind the times and make preparation for “innovative” teaching models for now and the future.

Constantly renewed and diversified technologies greatly change and innovate the teaching scenes. In addition, data collecting and analysis are considered more and more important with the revolutionary development of information technology and that of the Internet. Big data is also called 巨量資料 in Chinese. Fields including astronomy, atmospheric sciences, medicine, social networking, police administration and security, traffic information, and e-commerce all make use of data analysis to improve their work effectiveness or create business opportunities (Wikipedia, 2017). Education nowadays also goes with the trend to facilitate teachers’ professional skills using of big data collected from teaching scenes. Take National Education Technology Plan of the USA, 2016, for example, it covers concepts akin to AltSchool: use non-intruding, real-time and insensible ways that proceed along with daily activities to collect students’ learning data. Big data has changes finance, health, consumer technology retailing, and professional sports and so on. Along with other industries, the field of education has made systematical changes by using big data (e-paper of Ministry of Education, 2016). Faced with 12-Year Basic Education Curricula which is relatively new, it is an important question that how can teachers use teaching behavioral data to flip learning and teaching and meanwhile improve their professional skills.

II. Collection and Analysis of Teaching Behavioral Data

i The Washington-based Data Quality Campaign

The Washington-based Data Quality Campaign, The USA (DQC) proposed requirements of facilitating teachers’ data intelligence with regard to teachers’ professional development. In 2014, there are 19 states in the USA making data intelligence one of the requirements of teacher certification (Herold, 2014, cited Wang Ping, 2015). DQC has been implemented since 2005, with the view of encouraging and supporting policy makers from different states to improve their way of collecting, assessing, and using high quality education data. DQC, meanwhile, proposed 10 state Actions to make sure each state capitalizes on longitudinal data systems to constantly improving education. One of the actions is profession development: conduct policies and improve practices like profession development and certifications to ensure educators understand, properly assess, analyze and utilize data (Wikipedia, 2017). As a conclusion of the aforementioned, collecting, analyzing and utilizing education data are ways of improving the quality of education and necessary abilities teachers should have to refine teaching and learning.

ii. Teaching Strategies Based on Data and Personalized Learning

The biggest difference between 12-Year Basic Education Curricula and 9-Year Basic Education Curricula is that the former educates students in different ways according to students’ differences instead of having students experience the same learning process (Peng, Lien-Yi, 2017). In learning scenes, teachers should conduct differentiated instruction in accordance with students’ differences. With the use of data, learning is individualized, refined, and felt, taking place outside of the classroom (Chart 1). At this, the key points are how we “assess or collect” the “data” of teaching and learning. How can teachers collect real-time data to conduct teaching strategies and diversified teaching? How can students suggest the level of difficulty in an on-going lesson? How do students review after class? Most importantly, how do teachers improve their professional skills in teaching basing on teaching behavioral data after lessons?

Resources:Data Quality Campaign (2017). Chart 1 Data and Individualized Learning

iii. Sokrates Data Analysis System

The frame diagram of “Teaching behavioral data collection” function of Sokrates Data Analysis System is shown in Chart 2. Sokrates Data Analysis System automatically collects and analyzes teaching behavioral data. The data are as follow: 1. Activity Tendency Data: it shows if interaction in class is dominated by teachers or students. 2. Effective Interaction Data: it shows frequency of effective interaction in class and yields effective interaction points including those of the whole class and each group after computing the interaction behavioral data between students or between teachers and students. 3. Facility Usage Data: it shows the usage of devices in a Smart Classroom during the class, suggests frequency statistics of device usage, time rate and interaction intensity when individual devices are used. 4. Teaching Methods Application Data: The data shows the tendency of teaching methods applied in class as well as presents the traits of teaching methods applied in class after it analyzes teaching behavioral data according to teaching clips. 5. Questioning Label Data: the data suggests effective questioning label certified by experts using Sokrates Class Discussion Interface to complement and revise the deficiency of AI. 6. Response Label Data: the data suggests effective questioning label certified by experts using Sokrates Class Discussion Interface to complement and fix the defects of AI. Teachers’ teaching behaviors in class will be automatically recorded and uploaded to cloud platform by HiTeach Interactive Teaching System, then Teaching Behavior Big Data Cloud Service will automatically generates “Teaching Behavioral Data Analysis Report” (Chart 3).

Chart 2 Frame Diagram of Teaching Behavioral Data “Collection”
Sokrates Data Analysis System helps experts and auditing teachers conduct class discussion in a more scientific and effective way (Chang, Yi-Hua, Wu, Chuang-Wei, 2017). Massive teaching and learning big data collected and objectively analyzed by Sokrates Analysis System along with professional knowledge from teaching experts give instructions and suggestions on teachers’ teaching models, methods and many other aspects. Such practice reduces subjective factors that would affect teachers’ evaluation in teaching research. Its immediacy also promotes the efficiency on instructing teachers with their professional skills. Teachers may contrast their teaching design through teaching behavioral data analysis report generated after each class. This way, teachers may refine their teaching contents and ensure their class meets teaching aims, which is known as the autonomous mode of teachers’ selves improving on professional skills. Data Analysis helps teachers develop student-centered teaching; among Sokrates Analysis Statistics, teaching behavioral data analysis adds up and digitalizes teaching methods used by teachers in class to help find out strategies that best promote teaching effectiveness. Most importantly, all questions and response between teachers and students in class can be effectively recorded, which helps better interaction between teachers and students.

Chart 3 Teaching Behavioral Data Analysis Report Resources:Chang Yi-Hua, Wu Chuang-Wei2017)。

III. Conclusion

After observing and filming in class, experts in the past recorded what teachers teach with handmade forms and protocols and brought back the contents to analyze them in laboratories. With the technological advance in recent years, experts begin to record and analyze what teachers teach in class more effectively through technology. In contrast to traditional teaching analysis such as FIAS Interaction Analysis System (Flanders, 1970), S-T Analysis (藤田廣一、吉本英夫,1980), Seating Chart Observation Records (Gall & Acheson, 1980). Modern teaching analysis such as Communities of Practice, COP (Wang Lu, 2012), Teacher-Student Classroom Interactions(Hershkovitz, Merceron, & Shamaly, 2015), Wearable Sensors (Prieto, Sharma, Dillenbourg, & Jesús, 2016), analysis methods with mark systems and so on. These analysis systems do help refine teaching. However, there are many limitations and difficulties, including high HR cost. Several systems require many a workers to record in teaching scenes. Even though pen-paper recording has switched to computer assisted recording, the cost of human resources remains high; the results of analysis are difficult to understand and the explanation relies on experts. The resources of data are indirect. While some systems are established on online teaching platforms, data collected are students’ learning results or merely verbal interaction between teachers and students instead of first-hand data of the class. Data collecting was not easy, requiring additional gears. Some systems require teachers or experts to carry or wear additional devices. Thanks to ever-renewing technologies, “AI” is used to develop smart machines. “Smart” means that machines can react properly to the environments, predict and prepare for future events in certain conditions(Nilsson, 2009). Through smart teaching data collecting systems, we can break limitations and fix problems mentioned above, collect teachers’ behavioral data, analyze teaching behavioral traits and store class discussion records easily. With the assistance of “Sokrates Analysis System”, we collect teachers’ teaching behavioral traits. Using “Sokrates Analysis System”to collect teaching messages and conduct data analysis, teachers can transform the lesson. The system even helps teachers improve their professional skills so that teachers can realize smart education which give students diversified resources in accordance with their aptitude.


  • 王陸(2012)。教師線上實踐社區COP的績效評估方法與技術。中國電化教育,1,61-72。
  • 王萍(2015)。大數據時代提升教師數據智慧研究。開放教育研究,3,30-39。
  • 吳耀明(2010)。未來學校理想教育發展趨勢:以焦點訪談分析為例之建構。臺中教育大學,24(2),1-21。
  • 教育部電子報 (2016)。大數據分析的未來 : K-12教育 。取自
  • 林佳賢、陳昱儒(2014),〈青少年華語學習者線上互動策略研究〉,第13屆台華會年會暨國際學術研討會論文集,163頁。
  • 張奕華、吳權威(2017)。大數據趨勢下的教師專業發展平台:「醍摩豆智慧型議課廳」之發展與應用。輯於中華民國學校建築研究學會主編,學校建築與課程發展(168-181頁)。臺北市:中華民國學校建築研究學會。
  • 維基百科(2017) 。巨量資料。 取自
  • 維基百科(2018)。翻轉課堂。 取自翻转课堂
  • 藤田廣一、吉本英夫 (1980)。Binary表示による教師教育のための授業分析-S-T授業分析-。日本教育工学雑誌,5(3),119-128。
  • Brooks-Young, S.(2006). Critical technology issues for school leaders. Thousand Oaks, CA: Corwin.
  • Data Quality Campaign (2017). You Need Data to Personalize Learning. Retrieved from 17/08/DQC-You-Need-Data-to-Personalize-Learning.pdf